臺灣博碩士論文加值系統

高畫質(High definition)攝影機被廣泛使用在監視系統上，而焦距式高畫質攝影機可以用來監視一個廣大的區域，卻不方便以人工手動的方式長時間控制攝影機的光學變焦，因此需要設計自動化控制焦距的機制，才能拓展其應用領域。有鑑於此，本文提出一個針對固定式高畫質攝影機所設計的自動化焦距縮放機制，當少量的移動物體行經攝影機的可視範圍(Field of View)，會自動地控制攝影機的焦距，並盡可能地擷取物體的清晰影像，以提供有用的影像進行相關的影像服務，如：人臉辨識。為了達到上述的目標，這裡利用高斯混合模型(Gaussian Mixture Model)、連續影像相減法、CamShift追蹤方法、卡爾曼濾波器來偵測移動物體與追蹤，再利用自適應類神經模糊推理系統(Adaptive Neuro-Fuzzy Inference System)學習每次移動物體所調整的焦距值，用來估算適合的焦距調整值。依據雛形系統所進行的實驗，結果顯示本論文提出的自動化焦距縮放機制可以在實際的環境中擷取移動物體的清晰影像，而透過人臉偵測的結果也顯示出擷取的清晰影像的實用性。

High definition (HD) camera is widely used in surveillance systems. An HD camera with optical zoom is useful for monitoring a large area. However, it is inconvenient for a user to manually control the optical zoom for a long time. To exploit the functionality and extend the application domains of a HD camera, the zooming should be controlled automatically. Therefore, an automatic zooming mechanism is proposed in this paper. When the number of an object is small in the field of view (FOV) of the camera and an object is moving through the FOV, the zoom is controlled for capturing the object as clear as possible. A clear object image is useful for related image-based services, such as face recognition. In order to achieve the above goal, a Gaussian Mixture Model (GMM), temporal image differencing, a CamShift tracking method, and a Kalman filter are utilized for object detection and tracking. Then, an adaptive neuro-fuzzy inference system (ANFIS) is used to learn and determine a suitable value for adjusting the zoom. According to the experimental study of the prototype, the results show that the proposed mechanism is useful to capture the clear images of moving objects in a practical environment. A face detection algorithm is also used to demonstrate the feasibility of the captured clear images.

中文摘要 I
Abstract II
誌謝 III
目錄 IV
表目錄 VI
圖目錄 VII
第一章 簡介 1
第二章 文獻探討 3
第三章 系統設計 6
3.1 物體偵測模組 7
3.2 視覺追蹤模組 8
3.3 焦距控制階段 10
第四章 實驗結果與分析 15
4.1 實驗場景設置 15
4.2 實驗結果 16
第五章 結論 22
參考文獻 23
附錄一：3輸入的ANFIS訓練資料(上) 24
附錄二：3輸入的ANFIS訓練資料(下) 26
附錄三：3輸入的ANFIS訓練資料(左) 28
附錄四：3輸入的ANFIS訓練資料(右) 30
附錄五：3輸入的歸屬函數與控制規則(上) 32
附錄六：3輸入的歸屬函數與控制規則(下) 35
附錄七：3輸入的歸屬函數與控制規則(左) 37
附錄八：3輸入的歸屬函數與控制規則(右) 39

[1] S. C. Chan, S. Zhang, J. F. Wu, H. J. Tan, J. Q. Ni, and Y. S. Hung, “On the Hardware/Software Design and Implementation of a High Definition Multiview Video Surveillance System,” IEEE Journal on Emerging and Selected Topics in Circuits and Systems, Vol. 3, No. 2, June 2013, pp. 248-262.
[2] G. Scotti, A. Cuocolo, C. Coelho, and L. Marchesotti, “A Novel Pedestrian Classification Algorithm for a High Definition Dual Camera 360 Degrees Surveillance System,” IEEE International Conference on Image Processing (ICIP 2005), Genoa, Italy, 11-14 Sept. 2005, Vol. III, pp. 880-883.
[3] M. S. Sayed and J. G. R. Delva, “An Efficient Intensity Correction Algorithm for High Definition Video Surveillance Applications,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 21, No. 11, Nov. 2011, pp. 1622-1630.
[4] J. Chen, K. Benzeroual, and R. S. Allison, “Calibration for High-Definition Camera Rigs with Marker Chessboard,” 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Providence, Rhode Island, USA, 16-21 June, 2012, pp. 29-36.
[5] T. Nagai, T. Toyota, and T. Nagoya, “Implementation of High-Definition Lecture Recording System for Daily Use,” 2013 IEEE Global Engineering Education Conference, Berlin, Germany, 13-15 March, 2013, pp. 520-525.
[6] S. C. Jeng, “A GMM-based Method for Dynamic Background Image Model Construction with Shadow Removal,” Master Thesis, National Chiao-Tung University, ECE, June, 2005, pp. 83.
[7] G. R. Bradski, “Real Time Face and Object Tracking as a Component of a Perceptual User Interface,” in Proceedings of the Fourth IEEE Workshop on Applications of Computer Vision (WACV’98), 1998, pp. 214-219.
[8] G. Welch and G. Bishop. “An Introduction to the Kalman Flter,” University of North Carolina, Chapel Hill, Technical Report, TR95-041, 2004, 16 pages.
[9] J. S. R. Jang, “ANFIS: Adaptive-Network-Based Fuzzy Inference System,” IEEE Transactions on Systems, Man and Cybernetics, Vol. 23, 1993, pp. 665-685.